 Microsoft rules...
started 11/8/2002; 8:31:27 AM - last post 11/13/2002; 8:29:26 AM
|
|
Ehud Lamm - Microsoft rules... 
11/8/2002; 8:31:27 AM (reads: 1745, responses: 17)
|
|
| Microsoft rules... |
|
Bill Gates is expected to make some interesting announcements at OOPSLA:
Everett (the Visual Studio.Net update) will include a C++ compiler that is for the first time largely compliant with a standards set down by the International Organization for Standardization (ISO) nearly six years ago. Microsoft said its C++ compiler had been around 90 percent compliant with the standard. The Everett release will bring that closer to 98 percent, said Nick Hodapp, a Microsoft product manager.
Hodapp said the standards compliance is significant because it will now allow Visual C++ developers to take advantage of several popular development libraries, including Boost, Blitz and Loki. Libraries include pre-built code that let developers more quickly construct software applications...
That release of Visual C# will include four new features: support for "generics," which is a form of a C++ template that can help C# developers build software more quickly; support for "iterators," which help developers create new code; anonymous methods, which ease development of what’s known as "event-driven" code; and support for "partial types," which make it easier to use C# for building large projects.
So you see, there is progress...
Also: InfoWorld report.
Posted to general by Ehud Lamm on 11/8/02; 8:34:04 AM
|
|
|
|
Sam Gentile - Re: Microsoft rules...
11/8/2002; 10:45:52 AM (reads: 1430, responses: 0)
|
|
|
I am doing totally C++ stuff at the moment. I can confirm the compliance - partial template specializations, etc. In fact, with 7.1, I *am* building Loki, Boost and others without a single issue or problem today. Just built both of them today....
|
|
Patrick Logan - Re: Microsoft rules...
11/8/2002; 4:40:37 PM (reads: 1397, responses: 0)
|
|
|
Bringing C++ up to date is a good thing. Not sure it's worthy of an OOPSLA announcement. But maybe that reflects more on how irrelevant OOPSLA is compared to 10-15 years ago.
C# improvements... the challenge will be as with Java, only worse. Taking a fairly complicated language, adding more complications, and hoping the result is more usable.
The real excitement this week, as reported earlier in LtU, is the Lightweight Languages Workshop.
|
|
Sam Gentile - Re: Microsoft rules...
11/9/2002; 4:08:37 AM (reads: 1347, responses: 0)
|
|
|
Hi Patrick! Long time no see! Where have you been (no web log since August?)
I can't judge the caliber of OOPSLA announcements but what has been done in Everett C++ is truly fanstastic and reflects perhaps one of the most standards compliant C++ compilers available anywhere. I can't say much more but haing Herb Sutter, Stan Lippman and others on board have truly pushed this compiler to the forefront of the field.
|
|
Patrick Logan - Re: Microsoft rules...
11/9/2002; 7:23:47 AM (reads: 1324, responses: 0)
|
|
|
(Hi Sam. I'm on a weblog hiatus, not really intentionally, it just kind of evolved and I'm not ready to get back to it yet. Something like writers block. More like RSS feeds gave me "programmers block". I needed to stop soaking up ideas, and just write some code.)
Although I'm not in the C++ world, I appreciate what Microsoft is doing. Especially with their managed aspects, joining some of the earlier pioneers in bringing GC to C++.
I do think this reflects more on OOPSLA... it's just not cutting edge *research* that I can see. Good product engineering? Definitely.
|
|
Luke Gorrie - Re: Microsoft rules...
11/11/2002; 5:22:59 AM (reads: 1198, responses: 1)
|
|
|
As a C++ outsider, it seems pretty scary that it's a great achievement for a highly funded group to become "largely compliant" with a specification ~6 years after its publication. I guess it's at least proof that people don't listen to Niklaus Wirth and Tony Hoare.
(Also a little bit ironic since Tony Hoare works at Microsoft, and the Microsoft Research website seems happy to boast about the wisdom he brings to them.)
|
|
Ehud Lamm - Re: Microsoft rules...
11/11/2002; 5:42:40 AM (reads: 1230, responses: 0)
|
|
|
Sure. The Ada community really enjoys this sort of thing. Ada compilers ARE compliant.
But from a programmer's point of view, this is really good news. Now you'd be able to do all sorts of cool things with templates, that were impossible before. Though I hear most C++ programmers are fairly uneducated about templates in general.
|
|
Dan Shappir - Re: Microsoft rules...
11/11/2002; 6:00:53 AM (reads: 1176, responses: 0)
|
|
|
Most C++ programmers are scared shitless by templates, and the situation is even worse when it comes to exceptions (though Java is starting to change that). Still, library writers are generally not put off by templates, which means you get better libraries and thus better results (ATL is a case in point (when compared to the non-templated MFC)).
C++ is a very difficult language to be wholly compliant with. I think I remember hearing about one compiler that is almost wholly compliant, but that's about it. Case in point, the export keyword.
This is news not so much for the technical achievement (though it is quite an achievement if it's true) but for the change in direction which it signals. For years MS focus in VC++ was not C++ compliance. Rather it was to make VC++ a better tool for developing Win32 apps. Given the choice between adding another Wizard and fixing language compliance, the first choice always won.
When VC++ v7 came out I thought this trend was worsening. VC++ added a lot of new features that had nothing to do with C++. Instead, with managed code, it seemed to be moving away from the original language. I'm glad to see Microsoft doing the right thing. For me, at least, it makes a difference.
|
|
Sam Gentile - Re: Microsoft rules...
11/11/2002; 6:14:31 AM (reads: 1185, responses: 0)
|
|
|
>>This is news not so much for the technical achievement (though it is quite an achievement if it's true) but for the change in direction which it signals.
Bingo! That's whats important and it does matter to a lot of C++ programmers.
|
|
Sam Gentile - Re: Microsoft rules...
11/11/2002; 6:25:24 AM (reads: 1182, responses: 0)
|
|
|
>>>As a C++ outsider, it seems pretty scary that it's a great achievement for a highly funded group to become "largely compliant" with a specification ~6 years after its publication.
Luke, you seem hard to please-) This "well funded" group has done a lot more of course, but I am bound by NDA until release to not mention the rest of the things....FWIW
|
|
Luke Gorrie - Re: Microsoft rules...
11/11/2002; 9:00:19 AM (reads: 1164, responses: 0)
|
|
|
Yes, I'm a nitpicking bugger sometimes :-) and missed these strategic subtleties.
|
|
Michael Vanier - Re: Microsoft rules...
11/12/2002; 2:44:20 PM (reads: 1109, responses: 3)
|
|
|
Have Microsoft posted a specification about the enhancements to C#? I must say that I'm really surprised that they're adding such profound changes so quickly.
|
|
Ehud Lamm - Re: Microsoft rules...
11/12/2002; 2:50:08 PM (reads: 1128, responses: 0)
|
|
|
I was curious about that myself. Esp. the generics, of course...
|
|
Isaac Gouy - Re: Microsoft rules...
11/12/2002; 3:29:32 PM (reads: 1157, responses: 1)
|
|
|
|
Ehud Lamm - Re: Microsoft rules...
11/13/2002; 4:41:25 AM (reads: 1195, responses: 0)
|
|
|
Still, fi they are planning to introduce such features into the compiler, some specification of them must exist somewhere (but perhaps not publically accessible)
|
|
Noel Welsh - Re: Microsoft rules...
11/13/2002; 6:48:15 AM (reads: 1051, responses: 1)
|
|
|
What are "partial types"?
|
|
Ehud Lamm - Re: Microsoft rules...
11/13/2002; 6:59:06 AM (reads: 1088, responses: 0)
|
|
|
|
Patrick Logan - Re: Microsoft rules...
11/13/2002; 8:29:26 AM (reads: 1048, responses: 0)
|
|
|
|
|
|